For intermediate spot and futures traders, this guide gives exact crypto grid rules, sizing math, stop levels and failure conditions for range-bound markets.
Grid trading strategy crypto works when price keeps mean-reverting inside a defined range; it fails when you mistake a trend for chop. I use it as a rules-based range tool, not as a magic bot that fixes bad market selection.
The trader searching this is usually not asking what a limit order is. They want to know when a grid is worth running, how tight the levels should be, and how to stop one bad breakdown from wiping out 40 small wins.
A grid has edge when volatility is high enough to fill orders, but not so directional that price leaves the range and keeps going. BTC and ETH spot ranges on Binance or OKX are the cleanest place to start because spreads are tight and slippage is usually manageable.
My minimum filter is simple: two clean range reactions on both sides, no major news catalyst in the next 24 hours, and grid spacing at least 2.5x the round-trip fee. If the total fee is 0.20%, I want each completed grid to target at least 0.50% before slippage.
| Condition | Grid read | Action |
|---|---|---|
| BTC holds $58,000-$66,000 for 5-10 days | Range is tradable | Use spot grid or 1x-2x futures grid |
| Funding is above 0.10% per 8h on longs | Longs are crowded | Avoid long-biased futures grids |
| Open interest jumps 10% in 4 hours while price is flat | Breakout risk is rising | Reduce size or wait |
| Price closes 4H below range support twice | Range failed | Stop the bot and exit |
VoiceOfChain tracks range pressure, volume shifts and market sentiment in real time across Binance, Bybit and OKX - you can see when a grid is still trading chop versus when it is turning into a breakout risk. [voiceofchain.com]
Start with structure, not the bot preset. If BTC is trading near $62,000 and has rejected $66,000 twice while buyers defended $58,000 twice, the working range is $58,000-$66,000.
For that setup, 16 arithmetic grids create roughly $500 spacing. That is about 0.81% of BTC at $62,000, so after two 0.10% maker fees on Binance spot, the rough net per completed cycle is about 0.61% before slippage.
| Input | Example value | Why it matters |
|---|---|---|
| Current BTC price | $62,000 | Middle of the range is better than chasing the top |
| Lower grid | $58,000 | Below repeated support, not random |
| Upper grid | $66,000 | Near repeated rejection zone |
| Number of grids | 16 | Keeps spacing wider than fees |
| Approx interval | $500 / 0.81% | Enough room for profit after fees |
| Hard invalidation | $56,800 | About 2% below range support |
On Coinbase Advanced, I would only do this manually with post-only limit orders unless my fee tier is low, because maker and taker fees can be much larger than on major offshore venues. On KuCoin or Gate.io alt pairs, I widen spacing to 1.2%-2.5% because thin books can turn a clean grid into slippage bleed.
For futures grids on Bybit, OKX or Binance, I use isolated margin and low leverage. A 3x futures grid can look efficient until one liquidation cascade turns a normal 4% range break into a forced exit.
The clean rule is to size the grid from invalidation, not from how much capital is sitting idle. On a $10,000 account, I do not want one grid loss to exceed 1%-1.5%, so the planned loss should be about $100-$150.
Assume you deploy $3,000 into the BTC spot grid. If the bot ends up 80% allocated to BTC near the bottom, that is about $2,400 of BTC exposure; if your average cost is $60,000 and the stop is $56,800, the mark-to-market loss is about $128 before fees.
| Account size | $10,000 | Base account |
|---|---|---|
| Max risk | 1.25% | $125 planned loss |
| Grid capital | $3,000 | 30% of account |
| Estimated max BTC exposure | $2,400 | 80% filled near range low |
| Average BTC cost | $60,000 | Approx blended fill |
| Stop price | $56,800 | 5.33% below average cost |
| Estimated loss | $128 | Slightly above target before fees |
If that risk is too high, do not move the stop lower just to feel safer. Reduce the grid to about $2,300, widen the range, or skip the trade.
The most common mistake is running a grid as a disguised DCA strategy. A bot buying every 1% down is not market making if the coin is in a real downtrend; it is just averaging into weakness.
My hard risk caveat: grid trading fails hardest during regime shifts. If BTC breaks a 10-day range during a macro event, CPI release, ETF flow shock or liquidation cascade, the right trade is often to stop trading the range, not widen it.
The key takeaway: a crypto grid is a range strategy with inventory risk, not passive income. Use it when the market is liquid, bounded and volatile enough to pay after fees.
The setup is simple: define the range, make spacing wider than costs, size from the stop, and exit when the market stops behaving like a range. If you cannot name the invalidation level before launching the bot, the grid is not ready.